Correlation between venous pressure gradients and intravascular ultrasound in the diagnosis of iliac vein compression syndrome

Abstract

ObjectiveThe objective of this study was to evaluate the correlation between venous pressure gradients (VPGs) and intravascular ultrasound (IVUS) for the diagnosis of caval-iliac venous obstructions in patients with advanced chronic venous insufficiency. MethodsFifty patients with advanced chronic venous insufficiency symptoms (Clinical, Etiology, Anatomy, and Pathophysiology class 3 to 6) were prospectively submitted to multiplanar venography (MV) with intravenous pressure measurements and IVUS. The patients' lower limbs were divided accordingly: group I, limbs with <50% obstruction on IVUS (n = 49); and group II, limbs with ≥50% obstruction on IVUS (n = 51). Receiver operating characteristic curves compared the diagnostic performance of the VPGs. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy assessed the performance of VPGs in categories to determine the presence of significant obstruction. Logistic regression assessed the capacity of the VPGs to identify significant obstruction. ResultsThe most frequent point of venous compression according to IVUS was the proximal left common iliac vein (70%; P < .05). Group II showed a greater prevalence of transpelvic (group I, 8.2%; group II, 74.5%; P < .001) and paravertebral collaterals (group I, 4.1%; group II, 45.1%; P < .001) on MV. The femoral vein pressures at rest and after reactive hyperemia as well as the femorocaval gradient after reactive hyperemia (FCG-rh) and the femoral gradient after reactive hyperemia were significantly higher in group II (P = .001, P < .001, P = .002, and P = .006). The FCG-rh and the femoral gradient after reactive hyperemia presented the best diagnostic performance among the VPGs (P = .004 and P = .007) in the receiver operating characteristic curve analysis, although no significant differences between them were found. All the gradients presented low values of sensitivity (<40%), negative predictive value (<60%), and accuracy (<30%). Logistic regression showed that FCG-rh was significantly independent of MV (OR, 8.1; P = .011) in identifying significant obstructions. ConclusionsThere is correlation between the VPGs and significant obstructions with IVUS. However, this correlation does not translate to a good diagnostic performance of these VPGs. Only the FCG-rh added significant information to MV in identifying significant caval-iliac vein obstructions.